Approximately 30% of American adults suffer some kind of chronic pain. This profound public health issue is compounded by the serious side effects and abuse potential of the opiate painkillers (analgesics) used to help manage chronic pain. Thus new kinds of analgesics are desperately needed. Recently, an entire new class of analgesics (NK1-receptor antagonists) turned out to be ineffective in humans despite widespread success in rodents. These failures are part of a wider trend - FDA data shows that 85% of analgesics fail in human trials, while the average failure rate of all new drugs is roughly 89%. Few of these failures are due to side effects or toxicity. Instead in about 80% of cases, they simply don't work. These failures represent roughly 75% of the cost of drug development, creating financial pressures that drive widespread negative impacts on society (including the cost of drugs, and the lack of drug development for unprofitable diseases or populations). Recent failed analgesics ('pain killers') provide useful lessons for solving this broader problem. Every drug that fails in human trials appeared to work in animals - simply put; analgesic drug discovery in animals is very prone to false positives. One reason for this is that the measures of pain in mice which are used to discover new analgesics bear little correspondence to human clinical symptoms. For instance in mice we measure the suppression of reflex responses rather than the suppression of the sensation of pain itself; but in humans, we want to find drugs that blunt the sensation of pain and leave essential injury-preventing reflexes intact. Thus, in mice we are actually measuring exactly the wrong things if we want to discover a useful analgesic in humans. To solve this problem we ideally need to ask a mouse to rate the severity and type of pain it is experiencing in exactly the same way that we ask humans in a hospital recovery suite, but without the benefit of language. In fact, we already covertly monitor a human patient's pain in a way that is easily converted for use in mice - by giving the patient control over their own analgesia, and observing their self-dosing behavior. In humans this technique is called 'Patient Controlled Analgesia'. Animals can also selectively consume analgesics to control pain - a phenomenon called 'Analgesic Self Administration'. However, current Analgesic Self Administration techniques in animals are crude and poorly suited to drug discovery and translation. Therefore, this project will refine these techniques to develop a mouse equivalent of Patient Controlled Analgesia, so that pain in mice can be measured in the same way it is in humans. We will develop equipment that allows the use of this technique in the home cage of group-housed mice. We will test whether it can ask mice to report the severity and type of pain they are in, and whether it would have correctly identified as failures the very drugs which recently failed in humans. The resulting technology will be made publically available as open source software and 3D printing files, so that any researcher can build this new equipment. The impacts will be profound: for researchers we will deliver a translational predictive measure of pain in the mouse; and for patients this will mean the faster delivery of more effective and cheaper drugs.